I have a network with some hidden variables, for which I successfully trained the parameters (and thanks again for the great support!). Now I want to calculate a set of posteriors and likelihoods for the hidden variables over some additional data set of similar variables. I can do this manual (set evidence, update, read posteriors) but I have many vectors, and I need to do this automatically.
From my experience with research in BN, this is a common task. Once you have a network, you really want to use it to ask some questions on your data, otherwise whats the point?
Anyway, I would be delighted for any pointers.
Thanks!